Relation Between Language and False Belief Performance in High Functioning English Speaking Children with Autism

Background:

False belief tasks are one method of assessing Theory of Mind. The role of language in false belief has been studied; in particular English complementary structure and vocabulary understanding have been evaluated (deVilliers, 2005; Tager-Flusberg & Joseph, 2005).

Objectives: This study attempted to examine which aspect of language might predict false belief task performance in children with high functioning autism. The predictors examined were chronological age, comprehension of: mental verb (Think) and communication verb (Say), and relative clause, and receptive vocabulary. This was a part of larger study  funded by the Korea Research Foundation. The larger study is a cross-linguistic comparison between English- and Korean-speaking children.
Methods: Twenty children with high functioning autism (age between 5 and 11 years) and 20 children with language delay without autism (age between 3 and 6 years) will participate in this study. Data collection has completed so far with sixteen English speaking children with high functioning autism (Mean=7.6 years). Each participant received an IQ test (K-ABC-II) and a receptive vocabulary test (PPVT-4). All participants had IQ scores higher than 70. The ToM task consisted of unexpected location (FBL), unexpected content (FBC), language task (Say and Think) tasks. These were presented on a laptop computer while the examiner presented brief stories that corresponded to the pictures and asked questions.

Results: Preliminary data from the first 16 participants with autism was analyzed using multiple regression analyses.  After controlling for age (R²=.44, F (1, 14)=10.87, p=.005), the regression equation with the Think was significant, R² change=.22, F (1, 13)=8.51, p=.012. However, the regression equation was not significant with either Say, R2 change=.00, F (1, 12)=.02, p=.897 or Relative Clause (RC) comprehension, R² change=.03, F (1, 11)=1.12, p=.312. Also the regression equation with the PPVT was significant, R² change=.13, F (1, 10)=7.44, p=.021.
Table 1. Regression analysis using language as predictors.

step	variable	B	SE	b
1	age	1.06	.32	.66**
2	Age Think	.72 1.23	.28 .42	.45* .52*
3	Age Think Say	.71 1.20 .06	.30 .52 .51	.45* .50* .03
4	Age Think Say RC	.71 1.18 .01 2.33	.30 .51 .51 2.2	.44* .49* .01 .18
5	Age Think Say RC PPVT	.78 1.20 -.51 .10 .08	.24 .41 .45 1.93 .03	.49** .50* -.22 .008 .46*

Step 1, R²=.44, F (1, 14)=10.87, p=.005

Step 2, R²=.66, DR²=.22, F (1, 13)=8.51, p=.012

Step 3, R²=.66, DR²=.00, F (1, 12)=.02, p=.897

Step 4, R²=.69, DR²=.03, F (1, 11)=1.12, p=.312

Step 5, R2=.82, DR²=.13, F (1, 10)=7.44, p=.021

Conclusions: The preliminary results indicate that age, mental verb (Think) comprehension, and receptive vocabulary predicted the false belief performance. Results of this study are consistent with de Villiers and Pyers (2002) who proposed that the complement structure comprehension is critical in understanding false belief tasks.